Bruno Crispo / Research
Research
Research interests
Large-scale analysis of web-based attacks
Discovering new web vulnerabilities and measuring how many real websites are exposed to them. Results include the Relative Path Overwrite (RPO) attack and the Web Cache Deception work presented at USENIX Security 2020.
Security of the Internet of Things
Protecting IoT and cyber-physical systems against attack vectors that appear when subsystems never meant to be online are interconnected. The focus is on the simplest and most widely deployed devices, which have no special hardware for key protection and sometimes not even basic memory protection, so the emphasis is on software-based defences.
Behavioural biometrics
Authentication for smartphones, smart watches and fitness bands, where PINs and passwords are unpopular. Methods built on how users perform activities, using the MEMS sensors already in these devices, aiming at accuracy that users find easy to accept. The topic spans security and privacy, machine learning and human-computer interaction.
Adversarial learning
Foundational work on designing learning algorithms that make adversarial attacks harder, rather than yet another application of adversarial learning. Carried out in collaboration with Prof. Bergadano of the University of Torino.
Mobile platforms security and privacy
Two threads: strengthening the security models and mechanisms of smartphone platforms and operating systems without hurting usability, and analysing mobile applications by combining static and dynamic techniques, including code loaded at runtime and constructs that hide their semantics, such as reflection.
Research projects
CROSSCON
An open, flexible, portable and vendor-independent IoT security stack that runs across very different edge devices, from bare-metal boards with a few KB of RAM to devices with a hardware root of trust and trusted execution environments, to give a consistent security baseline across a whole IoT system. The stack specifications are formally verified (project budget above €6M).
Safe and Secure Industrial Internet of Things (S2)
Foundations, tools and methods for building safer and more secure safety-critical cyber-physical systems: preventing cyber attacks, detecting them when prevention fails, and containing their effects. A consortium of established Italian research groups.
SPECTRO
Specialised education in cybersecurity and robotics: two double-degree master's programmes (120 ECTS) with a minor in innovation and entrepreneurship, delivered by 12 higher-education institutions from 7 countries, 2 SMEs, a research centre and EIT Digital.
DUCA — Data Usage Control
A platform-independent framework of security and privacy-enhancing solutions that lets European citizens and organisations control their data. Use cases: smart energy, usage control for big data and AI, and collaborative mobility.